| A thermal model of the NSTAR ion thruster was developed in this dissertation. This model consisted of two parts. One part of the model developed the equations and relationships to predict the power deposition into the ion thruster from the plasma. For the cathodes, a heat transfer equation was developed that took into account gray body radiation heat transfer, conduction heat transfer, convection heat transfer, ohmic heating of the cathode wall from the emission current, cooling from field-enhanced thermionic emission, heating from ionic recombination, and heating from backstreaming energetic electrons. The temperatures derived from this equation for a cathode compared well to previous experimentally determined temperatures. Between 19.2 and 22.5 W is predicted to heat an NSTAR ion thruster cathode for operation. This power is mostly lost to thermal inefficiencies in the cathode. The less conductive a cathode and the more thermally insulated the insert can become, the more efficient the hollow-cathode should be. Power deposition in the discharge chamber region was found by using known voltages and currents from experiments, and by predicting plasma parameters and all unknown voltages and currents. The plasma parameters were based on a thruster performance model and the currents were found using charge continuity principles. The most significant amount of heating is predicted to be from the primary electrons on the anode. At a 0.5 kW throttling point, the primary electron current is predicted to be 2.75 A with a corresponding power deposition of 91.9 W (50% of the total predicted power deposition on the thruster). The primary electron current for the 2.3 kW throttling point is 4.35 A with 134.0 W of power deposited (42% of total predicted thruster power deposition). The second part of the model used the thermal codes SINDA and TRASYS to predict the temperatures of the NSTAR ion thruster. The thermal codes were able to predict the temperatures in the discharge chamber region to within 10°C for many operation conditions. The difference between the power deposition needed by SINDA for a temperature fit and the predicted power deposition is within 6%. |
